The basic characteristics of the stellar populations in galaxies are determined by the history of star formation and by the initial mass spectrum with which the stars are formed. Many attempts have been made to model the history of star formation in galaxies using various quasi-theoretical descriptions of star formation, but star formation remains a very poorly understood process and no theoretical understanding with any real predictive power has yet been attained. However, recent work suggests the importance for star formation of some mechanisms that have so far not been very extensively studied, and I shall mention some of these possible processes here, undeveloped though these ideas still are at present.
First, I shall review briefly the observational evidence concerning the history of star formation in galaxies of different Hubble type. Most of the information we have for galaxies other than our own comes from their colors, which can be compared with the predictions of models constructed with various assumed histories of star formation. Ever since the work of Tinsley (1968), it has been clear that most of the observed variation of color with Hubble type can be understood in terms of differing time scales for star formation, the reddest galaxies (E and SO) having formed the bulk of their stars at a relatively early stage, while the much bluer Sc galaxies have formed their stars much more gradually and at a rate that has remained nearly constant up to the present time. Recent studies of normal elliptical galaxies, reviewed by Faber in this volume, confirm that they have no detectable present star formation.